A code generation method and system for power communication planning scenarios

CN122111444APending Publication Date: 2026-05-29STATE GRID ELECTRIC POWER ECONOMIC RES INST IN NORTHERN HEBEI TECH CO LTD +1
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
STATE GRID ELECTRIC POWER ECONOMIC RES INST IN NORTHERN HEBEI TECH CO LTD
Filing Date
2026-02-24
Publication Date
2026-05-29

AI Technical Summary

Technical Problem

In the current power communication planning process, code generation relies on manual writing, which leads to long development cycles, easy introduction of errors, and difficulty in quickly adapting to changes in planning requirements. Existing automated code generation models have high computational resource requirements and are difficult to deploy directly in power communication planning auxiliary systems.

Method used

By employing model distillation technology, knowledge from high-performance code generation models is transferred to lightweight code generation models. Through structured processing of power communication planning requirements information, lightweight code generation models are generated, adapting to changes in planning requirements and reducing computational resource requirements.

Benefits of technology

The lightweight code generation model enables rapid adaptation to planning changes in power communication planning assistance systems, reduces training costs, improves model deployment efficiency and generation quality, and ensures the stability and consistency of generated code.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122111444A_ABST
    Figure CN122111444A_ABST
Patent Text Reader

Abstract

The application discloses a code generation method and system for a power communication planning scene, relates to the technical fields of power communication planning and intelligent software, and comprises the following steps: acquiring power communication planning demand information, structurally processing the power communication planning demand information, and generating a planning demand representation; inputting the power communication planning demand information and the planning demand representation into a pre-established code generation model for training, so as to automatically generate planning-related codes according to planning demands, and output a high-performance code generation model with the maximum ability of generating planning-related codes as the target; based on the high-performance code generation model, performing distillation training on a pre-established lightweight code generation model to obtain a trained lightweight code generation model; and inputting the planning demand representation into the trained lightweight code generation model to output corresponding planning codes; and the application can improve the convergence speed and generation performance of the code generation model.
Need to check novelty before this filing date? Find Prior Art